A decentralized flat control system for intelligent building

A decentralized flat control system is introduced into the intelligent building, where each zone of the intelligent building autonomously makes decisions through information exchange with their adjacent regions in a large scale control network. We formulate the decentralized control as a constrained optimization and decompose it into decentralized subproblems, then convert the centralized optimization into multi-objective problem. A decentralized algorithm for minimizing the sum of all the functions is proposed. The convergence property of decentralized algorithm is theoretically analyzed which considers both convex and non-convex systems with constraints, and then validates through simulation results.